Switch cabinet operation hidden danger monitoring and early warning system and method based on nanoparticle detection

Abstract

This invention provides a monitoring and early warning system and method for potential operational hazards in switchgear based on nanoparticle detection, relating to the field of power equipment condition monitoring and fault diagnosis technology. The system and method include S1: real-time sampling of nanoparticles in the air inside the switchgear; acquiring particle size distribution parameters, quantity concentration parameters, and unit time change parameters of the nanoparticles using a multi-parameter synchronous acquisition method; and continuously recording these parameters using time-series sampling to obtain a raw data set of nanoparticle operational characteristics. By performing real-time multi-parameter synchronous acquisition of nanoparticles in the air inside the switchgear, combined with time-series sampling technology, comprehensive monitoring of the particle size distribution, quantity concentration, and changes in nanoparticles is achieved, providing high-quality raw data support for subsequent data processing.

Description

Technical Field

[0001] This invention relates to the field of power equipment condition monitoring and fault diagnosis technology, specifically to a switchgear operation hazard monitoring and early warning system and method based on nanoparticle detection. Background Technology

[0002] The field of power equipment condition monitoring and fault diagnosis technology mainly involves online or offline monitoring, data acquisition, feature analysis, and fault identification of the operating status of various primary and secondary equipment in power systems during operation, in order to achieve equipment health status assessment, potential fault identification, and operational risk early warning. Among them, the switchgear operation hazard monitoring and early warning system and method based on nanoparticle detection refers to a technical solution for monitoring the operating status of switchgear. By detecting, analyzing, and evaluating the characteristics of nanoparticles generated during the operation of switchgear, it is used to identify potential operational hazards inside the equipment and output early warning information. Its purpose is to achieve early perception of switchgear operation hazards, risk level determination, and early warning prompts, thereby providing technical support for the safe operation, condition-based maintenance, and fault prevention of power equipment.

[0003] While existing technologies have made some progress in power equipment condition monitoring, they still face several shortcomings in practical applications. First, current technologies typically rely on traditional electrical and non-electrical quantity data for monitoring equipment operation, failing to collect and analyze more subtle changes in real time, such as the dynamic changes in nanoparticle generation, leading to a lag in the monitoring system's response to minor faults. Second, most traditional solutions fail to fully consider the impact of environmental changes, such as temperature, humidity, and background particulate matter, in data processing. These factors can affect the accuracy and reliability of monitoring data, causing misjudgments or missed diagnoses. Finally, in fault identification and early warning, existing technologies mostly employ static models, lacking dynamic prediction of potential hazard development trends. This results in the inability to accurately assess the evolution path and risk level of potential hazards before equipment failure occurs, thus missing the optimal maintenance opportunity. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a monitoring and early warning system and method for potential operational hazards in switchgear based on nanoparticle detection. This solves the problem that existing technologies, in monitoring the operating status of equipment, typically rely on traditional electrical and non-electrical quantity data, which cannot collect and analyze more subtle changes in real time, such as the dynamic changes caused by the generation of nanoparticles, resulting in a lag in the response of the monitoring system to minor faults.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a monitoring and early warning system and method for potential operational hazards in switchgear based on nanoparticle detection, comprising the following steps:

[0006] S1: Real-time sampling of nanoparticles in the air inside the switch cabinet; using a multi-parameter synchronous acquisition method for nanoparticles to obtain the particle size distribution parameters, quantity concentration parameters, and unit time change parameters of nanoparticles; and continuously recording them through time series sampling to obtain the original operating characteristic dataset of nanoparticles.

[0007] S2: Based on the original operational feature dataset of the nanoparticles, the time series feature extraction method and the sliding time window analysis method are used to extract the trend features, fluctuation features and mutation features of the nanoparticle parameters, construct the time series feature expression of the nanoparticle operational state, and obtain the time series evolution feature parameter set of nanoparticles;

[0008] S3: Based on the nanoparticle time-series evolution characteristic parameter set, combined with the switchgear operating environment parameters, including temperature parameters, humidity parameters and background particulate matter concentration parameters, the environmental impact adaptive correction method is used to dynamically compensate and correct the nanoparticle characteristic parameters to obtain the nanoparticle electrical related correction characteristic parameter set;

[0009] S4: Based on the set of electrical-related modified feature parameters of the nanoparticles, the hidden danger pattern matching identification method is adopted to compare the modified feature parameters with the pre-established typical hidden danger feature model of the switchgear, and the operation hidden danger type and corresponding risk level are identified by feature similarity calculation and classification judgment method to obtain the switchgear operation hidden danger identification result set.

[0010] S5: Based on the identified results set of potential hazards in the switchgear, and combined with the change rate and evolution trend of the nanoparticle characteristic parameters, a hazard development trend prediction method is used to predict the direction of hazard evolution, and corresponding level hazard warning information is generated through a graded warning output method to obtain hazard warning information for switchgear operation.

[0011] Preferably, step S1 includes the following steps:

[0012] S101: Based on the electrical and thermal stresses generated during the operation of the switchgear, nanoparticle detection units are deployed inside the switchgear and along the ventilation path to sample nanoparticles in the air inside the cabinet in real time, obtain the particle size parameters, number concentration parameters, and distribution parameters of the nanoparticles, and obtain the original monitoring dataset of nanoparticles.

[0013] S102: Based on the original monitoring dataset of nanoparticles, the sampled data is filtered, denoised and smoothed using data preprocessing methods to eliminate random noise and unstable interference, and a preprocessed dataset of nanoparticles is obtained.

[0014] S103: Based on the nanoparticle preprocessing dataset, the nanoparticle parameters are continuously recorded and organized using a time series sampling method to form time-series data for subsequent analysis, thus obtaining the original nanoparticle operational characteristic dataset.

[0015] Preferably, step S2 includes the following steps:

[0016] S201: Based on the original operational feature dataset of nanoparticles obtained in step S1, the trend characteristics of nanoparticle size and number concentration changing with time are analyzed using the time series feature extraction method to obtain the time series trend feature dataset of nanoparticles.

[0017] S202: Based on the nanoparticle time-series trend feature dataset, the periodic fluctuations and abnormal fluctuations of nanoparticle parameters are extracted using the fluctuation feature analysis method to obtain the nanoparticle time-series fluctuation feature dataset.

[0018] S203: Based on the nanoparticle temporal fluctuation feature dataset, the sudden change feature of nanoparticle parameters is identified by the mutation point detection method, forming a complete temporal feature expression, and obtaining the nanoparticle temporal evolution feature parameter set.

[0019] Preferably, step S3 includes the following steps:

[0020] S301: Based on the nanoparticle time-series evolution characteristic parameter set obtained in step S2, temperature parameters, humidity parameters and background particulate matter concentration parameters in the switch cabinet operating environment are collected simultaneously to form an environmental impact parameter dataset;

[0021] S302: Based on the environmental impact parameter dataset, an environmental impact adaptive correction method is used to dynamically compensate and correct the nanoparticle characteristic parameters to obtain an environmentally corrected nanoparticle characteristic parameter set.

[0022] S303: Based on the environmentally corrected set of nanoparticle characteristic parameters, a correlation analysis method is used to screen characteristic parameters that are highly correlated with the electrical operating state, thereby obtaining a set of nanoparticle electrical correlation corrected characteristic parameters.

[0023] Preferably, step S4 includes the following steps:

[0024] S401: Based on the set of electrical-related modified feature parameters of nanoparticles obtained in step S3, the feature parameters are matched with the pre-established typical hidden danger feature model of switchgear using the hidden danger pattern matching identification method to obtain a preliminary result set of hidden danger type matching.

[0025] S402: Based on the preliminary result set of the hazard type matching, the feature similarity calculation method is used to accurately identify the hazard type, and a precise hazard type matching result set is obtained;

[0026] S403: Based on the accurate matching result set of the hazard types, the risk level determination method is used to determine the development stage and risk level of the hazard, and the result set of hazard identification for switchgear operation is obtained.

[0027] Preferably, step S5 includes the following steps:

[0028] S501: Based on the switchgear operation hazard identification result set obtained in step S4, and combined with the change rate and evolution trend of nanoparticle characteristic parameters, the hazard development trend prediction method is used to predict the evolution direction of the operation hazard, and the hazard development trend prediction result set is obtained.

[0029] S502: Based on the predicted result set of the hidden danger development trend, a hierarchical early warning output method is adopted to generate corresponding level of operation hidden danger early warning information according to the prediction results and preset thresholds, thereby obtaining switchgear operation hidden danger early warning information.

[0030] Preferably, the switchgear operation hazard monitoring and early warning system based on nanoparticle detection includes the following modules: a nanoparticle sensing module, which is used to perform multi-parameter real-time detection of nanoparticles in the air inside the switchgear based on the electrical stress and thermal stress generated during the operation of the switchgear, and to perform data preprocessing and time series construction on the obtained nanoparticle particle size parameters, quantity concentration parameters and change characteristics to generate a dataset of original nanoparticle operation characteristics.

[0031] The feature analysis and correction module is used to receive the original operating feature dataset of the nanoparticles, perform time series feature analysis on the nanoparticle parameters, extract the trend features, fluctuation features and mutation features, and combine the switch cabinet operating environment parameters to perform adaptive correction and correlation screening on the features, and generate a set of corrected electrical features parameters of nanoparticles.

[0032] The hazard warning and decision-making module is used to receive the set of electrical-related modified feature parameters of the nanoparticles, match and analyze them with the pre-built switchgear operation hazard feature model to identify the type and risk level of switchgear operation hazards, predict the development trend of operation hazards, output corresponding level operation hazard warning information, and generate switchgear operation hazard warning information.

[0033] Preferably, the nanoparticle sensing module includes:

[0034] The particle acquisition submodule is used to sample nanoparticles in the air inside the switchgear in real time based on the electrical and thermal stress generated during the operation of the switchgear, and to obtain the particle size parameters, number concentration parameters and distribution parameters of the nanoparticles.

[0035] The signal preprocessing submodule is used to filter, denoise, and smooth the raw sampling signal acquired by the particle acquisition submodule in order to eliminate the influence of environmental interference and random noise.

[0036] The temporal construction submodule is used to continuously sample and organize the preprocessed nanoparticle parameters to construct temporal nanoparticle operational status data.

[0037] The data integration submodule is used to fuse and encapsulate the time-seriesd nanoparticle operational status data to form the original operational feature dataset of the nanoparticles.

[0038] Preferably, the feature analysis and correction module includes:

[0039] The trend extraction submodule is used to extract the trend features of nanoparticle parameters changing over time based on the original operational feature dataset of the nanoparticles.

[0040] The fluctuation analysis submodule is used to analyze the periodic fluctuation characteristics and abnormal fluctuation characteristics of nanoparticle parameters.

[0041] The mutation identification submodule is used to detect sudden changes in nanoparticle parameters and identify abnormal mutation characteristics.

[0042] The environmental correction submodule is used to adaptively compensate and dynamically correct the nanoparticle characteristic parameters by combining the temperature parameters, humidity parameters and background particulate matter concentration parameters in the operating environment of the switchgear.

[0043] The relevant screening submodule is used to use correlation analysis methods to screen parameters that are highly correlated with the electrical operating state from the modified feature parameters, forming the set of electrical-related modified feature parameters of the nanoparticles.

[0044] Preferably, the hazard early warning decision module includes:

[0045] The hidden danger matching submodule is used to perform matching analysis between the feature parameters and the pre-built typical operation hidden danger feature model of switchgear based on the set of electrical-related modified feature parameters of the nanoparticles, so as to obtain the hidden danger type matching result;

[0046] The risk assessment submodule is used to determine the development stage and risk level of the switchgear operation hazards based on the hazard type matching results.

[0047] The trend prediction submodule is used to predict the development direction of potential operational hazards by combining the rate of change and evolution trend of the characteristic parameters of nanoparticles.

[0048] The early warning output submodule is used to generate corresponding level of operational hazard early warning information based on the risk level and prediction results, and output the switchgear operational hazard early warning information.

[0049] This invention provides a monitoring and early warning system and method for potential operational hazards in switchgear based on nanoparticle detection. It has the following beneficial effects:

[0050] This invention achieves comprehensive monitoring of nanoparticle size distribution, quantity concentration, and changes by real-time multi-parameter synchronous acquisition of nanoparticles in the air inside switchgear and combining it with time-series sampling technology. This provides high-quality raw data support for subsequent data processing. Based on this real-time acquired data, time-series feature extraction and sliding time window analysis methods can accurately extract the changing trends, fluctuations, and abrupt changes of nanoparticle parameters, helping to identify potential problems in switchgear operation at an early stage. Unlike traditional technologies, dynamic compensation and correction, combined with environmental factors such as temperature, humidity, and background particulate matter concentration, further improve the accuracy of nanoparticle data, reduce interference from environmental factors, and ensure the reliability and validity of the data. Furthermore, through hazard pattern matching identification and risk level determination, it can clearly identify the types of potential hazards in the switchgear and provide accurate risk level assessments. Combined with hazard development trend prediction, it can predict the development direction and potential dangers of hazards in real time, providing early warnings for operation, improving the system's intelligence and response speed, thereby effectively reducing the risk of equipment failure and ensuring the safe and stable operation of the power system. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of the main steps of the present invention;

[0052] Figure 2 This is a detailed schematic diagram of S1 of the present invention;

[0053] Figure 3 This is a detailed schematic diagram of S2 of the present invention;

[0054] Figure 4 This is a detailed schematic diagram of S3 of the present invention;

[0055] Figure 5 This is a detailed schematic diagram of S4 of the present invention;

[0056] Figure 6 This is a detailed schematic diagram of S5 of the present invention;

[0057] Figure 7 This is a system block diagram of the present invention. Detailed Implementation

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] Example:

[0060] like Figure 1-7 As shown, this invention provides a monitoring and early warning system and method for potential operational hazards in switchgear based on nanoparticle detection, comprising the following steps:

[0061] S1: Real-time sampling of nanoparticles in the air inside the switch cabinet; using a multi-parameter synchronous acquisition method for nanoparticles to obtain the particle size distribution parameters, quantity concentration parameters, and unit time change parameters of nanoparticles; and continuously recording them through time series sampling to obtain the original operating characteristic dataset of nanoparticles.

[0062] S2: Based on the original operational feature dataset of nanoparticles, time series feature extraction method and sliding time window analysis method are used to extract the changing trend features, fluctuation features and mutation features of nanoparticle parameters, construct the time series feature expression of nanoparticle operational state, and obtain the time series evolution feature parameter set of nanoparticles;

[0063] S3: Based on the time-series evolution characteristic parameter set of nanoparticles, combined with the operating environment parameters of the switchgear, including temperature parameters, humidity parameters and background particulate matter concentration parameters, an environmental impact adaptive correction method is used to dynamically compensate and correct the characteristic parameters of nanoparticles, and obtain the set of corrected electrical related characteristic parameters of nanoparticles.

[0064] S4: Based on the set of electrical-related modified feature parameters of nanoparticles, the hidden danger pattern matching identification method is adopted to compare the modified feature parameters with the pre-established typical hidden danger feature model of switchgear, and identify the type of operation hidden danger and corresponding risk level through feature similarity calculation and classification judgment method, so as to obtain the set of operation hidden danger identification results of switchgear;

[0065] S5: Based on the result set of switchgear operation hazard identification, combined with the change rate and evolution trend of nanoparticle characteristic parameters, the hazard development trend prediction method is used to predict the evolution direction of the hazard, and corresponding level operation hazard warning information is generated through hierarchical warning output to obtain switchgear operation hazard warning information.

[0066] S1 includes the following steps:

[0067] S101: Based on the electrical and thermal stresses generated during the operation of the switchgear, nanoparticle detection units are deployed inside the switchgear and along the ventilation path to sample nanoparticles in the air inside the cabinet in real time, obtain the particle size parameters, number concentration parameters, and distribution parameters of the nanoparticles, and obtain the original monitoring dataset of nanoparticles.

[0068] Based on the stress sources generated by the switchgear's long-term voltage loading and current changes during operation, the corresponding execution process revolves around the deployment and sampling of detection units. During execution, detection units are installed in the busbar connection area, near the contacts, and in the airflow channel, forming a multi-point sampling structure through these location differences. During sampling, instantaneous particle size distribution data and particle number data output by the particle counter are read according to a fixed sampling period. For example, the particle number changes at different locations are recorded within the same time slice. The particle size ranges at each sampling point are then categorized into small, medium, and large ranges, with the determination of the corresponding range based on... The response range given in the calibration document of the detection equipment is used to count the number of particles in different intervals within the same sampling period. The statistical process is to obtain the number concentration expression by summing the count values ​​within the sampling time period. The number concentration value is obtained by converting the cumulative number of particles per unit time to the sampling volume. The distribution parameter is calculated by the proportion of the number in each particle size interval. The proportion is expressed as the ratio of the number in a single interval to the total number. Relevant parameters are recorded synchronously during the sampling process to form multidimensional data entries. In the example, the sampling results within a certain time period can be organized into a data structure containing location identifiers, particle size interval identifiers, and corresponding number identifiers. Finally, the data is summarized to form a data set for subsequent processing.

[0069] S102: Based on the original monitoring dataset of nanoparticles, the sampled data is filtered, denoised and smoothed using data preprocessing methods to eliminate random noise and unstable interference, and a preprocessed dataset of nanoparticles is obtained.

[0070] Based on the established dataset, the data processing stage begins. The corresponding execution process revolves around signal quality management. During execution, the original sampled sequence is imported into the data processing environment. After uniform calibration of the time axis, filtering is performed. The filtering process is completed using a moving average method, which involves weighting the values ​​of several adjacent sampling points. The weighting coefficient is set according to the sampling stability, and the weight is set with reference to the sampling fluctuation range. When the difference between adjacent points is within a small range, a near-uniform weight is used; when the difference is within a large range, the weight of outliers is reduced. The weight values ​​are selected based on empirical intervals. The denoising process is implemented using a threshold determination method. The threshold is derived from the background noise statistics during the equipment's no-load operation. Data exceeding the upper limit of the background fluctuation range is marked as interference and replaced. The replacement value is represented by the mean of adjacent valid points. Smoothing is completed by performing a second window averaging on the processed sequence. The window width is determined based on the sampling frequency. In the example, the quantity data of several consecutive sampling times can be selected for calculation. The sudden increase in value at a single point is compared with the data before and after. When the difference exceeds the preset fluctuation range, it is corrected. All processing steps are consistent on the time axis. After processing, a consistent and continuous set of data entries is formed.

[0071] S103: Based on the nanoparticle preprocessing dataset, the nanoparticle parameters are continuously recorded and organized using a time series sampling method to form time-series data for subsequent analysis, thus obtaining the original nanoparticle operational characteristic dataset.

[0072] Based on the processed dataset, continuous recording and organization are carried out. The corresponding execution process revolves around time correlation. Data from each sampling moment is written into a time series data table in chronological order. During the recording process, a time index and a location index are added to each data point. The time index is incremented using a unified time base, and the location index is used to distinguish the sources of different detection points. During the processing, data from missing moments are interpolated using a linear estimation method based on the values ​​of adjacent moments. The estimation formula is calculated by the ratio of the difference between the values ​​of two adjacent points to the time interval. In the parameter description, the values ​​of adjacent points are expressed as the number of particles at the preceding and following moments, and the time interval is the sampling period length. During the processing, data from different locations at the same moment are also stored side by side to form a multi-dimensional time section record. In the example, the data within a certain continuous running segment can be arranged into a table format with time progression. Each row corresponds to a sampling moment, and each column corresponds to a particle size range or sampling location. After processing, the integrity of the time series is verified based on the continuity of the time index and the consistency of the data dimensions. Finally, a continuous time series dataset that can be directly used for subsequent analysis is obtained.

[0073] S2 includes the following steps:

[0074] S201: Based on the original operational feature dataset of nanoparticles obtained in step S1, the trend characteristics of nanoparticle size and number concentration changing with time are analyzed using the time series feature extraction method to obtain the time series trend feature dataset of nanoparticles.

[0075] Based on the original operational characteristic dataset of nanoparticles, the process first analyzes the trends of nanoparticle size and quantity concentration over time. First, particle size and quantity data for each moment in the time series are extracted. Then, the data points are arranged according to the time axis, and a moving average method is used to smooth the fluctuations. Using a specific time window as a unit, the maximum, minimum, and average particle size values ​​are selected and calculated within each window. By calculating the mean and variance of the particle size distribution, a trend data set is formed to further determine whether it shows a stable increase or periodic fluctuations. For example, if the average particle size increases significantly within a certain time period, and the quantity concentration also shows an upward trend, it indicates that there may be a gradually accumulating hidden danger in the equipment. By calculating the trend coefficient and rate of change, the speed of change is quantified to evaluate the trend characteristics of the system, finally obtaining the nanoparticle time-series trend characteristic dataset.

[0076] S202: Based on the nanoparticle time-series trend feature dataset, the fluctuation feature analysis method is used to extract the periodic fluctuations and abnormal fluctuations of nanoparticle parameters, and obtain the nanoparticle time-series fluctuation feature dataset.

[0077] Based on a time-series trend feature dataset of nanoparticles, the process first performs volatility analysis on the data to identify the system state by extracting periodic and anomalous fluctuations from the time-series data. Volatility analysis extracts the differences at each moment in the time series data, calculates the standard deviation and root mean square value for each time period, thereby measuring the range of fluctuations in particle quantity and analyzing whether the fluctuation frequency has significant periodicity. If the fluctuation frequency is significantly higher in a certain time period than in other time periods, it is considered an anomalous fluctuation. Using periodic analysis, the time-series data is transformed into the frequency domain using Fast Fourier Transform (FFT) to identify the frequency values ​​of periodic fluctuations. Assuming the frequency of periodic fluctuations in a certain time period is 0.2 Hz, adjustments are made based on this frequency to predict potential changes in the system. By calculating the ratio of the amplitude of periodic fluctuations to normal fluctuations, the presence of anomalous fluctuations is further determined, ultimately generating a volatility feature dataset.

[0078] S203: Based on the nanoparticle temporal fluctuation feature dataset, the sudden change feature of nanoparticle parameters is identified by the mutation point detection method, forming a complete temporal feature expression and obtaining the nanoparticle temporal evolution feature parameter set.

[0079] Based on a time-series fluctuation feature dataset of nanoparticles, a mutation point detection method is used for identification during execution. First, the rate of change of the fluctuation feature data at each time step is calculated by performing a difference operation. When a mutation point occurs, the rate of change increases significantly. Therefore, a mutation threshold can be set to determine whether the change of a data point exceeds the normal fluctuation range. Mutation point identification is achieved by calculating the slope and inflection points of the fluctuation sequence. If the fluctuation degree changes drastically at a certain time step, for example, suddenly increasing to more than three times the original fluctuation amplitude, it is marked as a mutation point. Continuous analysis is performed using a detection algorithm, and stable interval values ​​are obtained through interval division. If multiple consecutive mutation points are detected, it indicates that the changes within that time period are relatively drastic, possibly indicating an equipment anomaly or malfunction. By statistically analyzing the number and distribution of mutation points, a time-series evolution feature dataset is formed.

[0080] S3 includes the following steps:

[0081] S301: Based on the nanoparticle time-series evolution characteristic parameter set obtained in step S2, temperature parameters, humidity parameters and background particulate matter concentration parameters in the switch cabinet operating environment are collected simultaneously to form an environmental impact parameter dataset;

[0082] Based on the existing set of time-series evolution characteristic parameters of nanoparticles, the execution process revolves around the synchronous acquisition and processing of environmental parameters. First, environmental monitoring units are deployed inside the switchgear and its adjacent space to sample temperature, humidity, and background particulate matter levels in the air in parallel. The sampling process and nanoparticle characteristic records are aligned using a unified time stamp. Environmental parameters and nanoparticle parameters at the same moment are correlated through timestamp matching. Subsequently, the data of various environmental parameters are processed. Temperature parameters are divided into low, normal, and high ranges according to operating intervals; humidity parameters are labeled into dry, moderate, and humid ranges; and background particulate matter levels are divided into baseline and fluctuating ranges based on long-term statistical results. During parameter acquisition, environmental parameter values ​​are calculated from the raw signals output by the sensors. The background particulate matter concentration is obtained by converting the particle count per unit time to the sampling volume; the conversion relationship is expressed in symbolic form. ,in Indicates the background particulate matter level. This indicates the number of particles within the sampling period. The corresponding sampling volume is represented by the above steps, which form a data entry containing time index, environmental parameter category and corresponding value. This data is then integrated with the nanoparticle evolution characteristics to finally form an environmental impact parameter dataset.

[0083] S302: Based on the environmental impact parameter dataset, the environmental impact adaptive correction method is used to dynamically compensate and correct the nanoparticle characteristic parameters to obtain the environmentally corrected nanoparticle characteristic parameter set.

[0084] Based on the established environmental impact parameter dataset, the execution process revolves around dynamic compensation and correction. First, representative quantities from the nanoparticle characteristic parameters are selected as correction targets, such as the amplitude of particle size variation and the amplitude of quantity fluctuation. Then, the environmental parameters at the corresponding time points are incorporated into the correction calculation. Influence coefficients are set for temperature and humidity, with the coefficients set based on the statistical relationship between nanoparticle parameters and environmental changes over historical operating periods. When the temperature is in a higher range, a relatively larger correction coefficient is used; when it is in a lower range, a relatively smaller correction coefficient is used. The humidity correction coefficient is determined in a similar manner. The dynamic compensation process is expressed through a weighted correction formula, which can be represented in symbolic form as follows: ,in This represents the corrected characteristic parameters of the nanoparticles. Represents the original feature parameters. and These represent the temperature and humidity correction factors, respectively. and This represents the deviation between the current environmental parameters and the median of the baseline range. In actual calculations, parameters from a certain operating period can be selected for substitution and calculation. By comparing the changes in parameters before and after correction, it can be determined whether the correction is within a reasonable range. Then, the background particulate matter level is introduced as an additional correction term to readjust the data within the background fluctuation range. Finally, the set of nanoparticle characteristic parameters after environmental correction is obtained.

[0085] S303: Based on the environmentally corrected set of nanoparticle characteristic parameters, correlation analysis is used to screen characteristic parameters that are highly correlated with electrical operating conditions, thus obtaining the set of nanoparticle electrical correlation corrected characteristic parameters.

[0086] Based on the environmentally corrected set of nanoparticle characteristic parameters, the execution process revolves around feature selection and correlation calculation. First, the corrected characteristic parameters are time-aligned with the switchgear operating status parameters to form a multivariate comparison data table. Then, the correlation between each type of nanoparticle characteristic parameter and the operating status parameters is calculated. The correlation coefficient is used to describe the correlation, and its symbolic expression can be represented as follows: ,in This represents the value of a characteristic parameter of a certain nanoparticle at time i. This represents the running status parameters at the corresponding time point. and Each represents its average level. By dividing the correlation coefficient results into intervals, parameters with high correlation are marked as highly correlated features, parameters with medium correlation are marked as generally correlated features, and other parameters are not included in subsequent processing. During the screening process, the calculation results of multiple time periods are compared. When a feature maintains a high correlation interval in most time periods, it is retained. Through the above calculation and judgment steps, the feature parameters are screened and organized, and finally a set of nanoparticle electrical correlation correction feature parameters is formed.

[0087] S4 includes the following steps:

[0088] S401: Based on the set of electrical-related modified feature parameters of nanoparticles obtained in step S3, the feature parameters are matched with the pre-established typical hidden danger feature model of switchgear using the hidden danger pattern matching identification method to obtain a preliminary result set of hidden danger type matching.

[0089] Based on the existing set of electrical-related correction feature parameters for nanoparticles, the process first involves structuring various feature parameters, encoding particle size evolution characteristics, quantity fluctuation characteristics, and environmental correction correlation characteristics according to a unified parameter dimension. Then, a pre-established typical switchgear hazard feature model library is invoked, mapping the feature parameter vectors extracted during the current runtime to the corresponding feature dimensions in the model library. During the matching process, a matching weight is set for each feature, with the weight set in intervals based on the frequency and stability of the feature in historical hazard samples. The matching operation is completed by calculating the deviation between the current feature value and the model's baseline feature value. The deviation can be expressed as a difference ratio, and its symbolic form can be represented as... ,in Indicates the current running characteristic parameters. The baseline parameter represents the corresponding feature in the model. When the deviation falls into the preset low deviation range, it is judged as a high match; when it falls into the medium range, it is judged as a general match; and the rest are considered as a weak match. In a real-world scenario, data from a certain operating cycle can be selected for demonstration. The matching results of multiple features are summarized, and a comprehensive matching score is formed by weighted summation. The scores are then categorized according to the hazard model category corresponding to the score range, and finally, a preliminary result set of hazard type matching is obtained.

[0090] S402: Based on the preliminary result set of hazard type matching, the feature similarity calculation method is used to accurately identify the hazard type, and a precise hazard type matching result set is obtained;

[0091] Based on the preliminary result set of hazard type matching, the candidate hazard types are further refined during the execution process. First, hazard types with matching scores in the medium to high range are selected from the preliminary results as comparison targets. Then, for each candidate hazard type, its corresponding standard feature vector is extracted, and a dimension-wise similarity calculation is performed between this vector and the corrected feature vector under the current operating state. The similarity calculation is described using normalized correlation, and its expression can be represented as follows: ,in Indicates the current feature component, The similarity results obtained by the calculation are divided into pre-defined intervals, and the hazard types with high similarity are judged as highly similar, while those with medium similarity are judged as generally similar. During the execution process, the similarity results of different hazard types are compared, and the type with the highest similarity and the stability requirement is selected as the identification object. For example, if the same hazard type maintains a high similarity interval in multiple consecutive time windows within a certain running period, then the type is retained and the other types are eliminated. Through the above comparison and screening process, a set of accurate matching results for hazard types is formed.

[0092] S403: Based on the accurate matching result set of hazard types, the risk level determination method is used to determine the development stage and risk level of the hazard, and the result set of hazard identification for switchgear operation is obtained.

[0093] Based on the precise matching result set of hazard types, the development stage and risk level of the hazards are determined during the execution process. First, the corresponding risk assessment parameter table is retrieved for each identified hazard type. The magnitude, duration, and frequency of change of nanoparticle-related characteristics are used as input items. A risk weight is assigned to each input item, with the weight settings configured in intervals based on the characteristic performance of the hazard at different stages. Subsequently, each risk factor is quantitatively calculated. The risk calculation process can be represented using a weighted summation form, and its symbolic form can be expressed as follows: Where R represents the comprehensive risk index, This represents the weight coefficient of the j-th risk factor. The normalized parameter value of the corresponding risk factor is represented. The calculated comprehensive risk index is divided into different risk level intervals according to the preset interval. At the same time, the stage of the hidden danger is judged by the changing trend of the characteristic parameter. When the risk index is in the low interval and the change is slow, it is judged as the early stage. When it is in the medium interval and the change is continuous, it is judged as the development stage. When it is in the high interval and the change is obvious, it is judged as the aggravation stage. The stability of the judgment is confirmed by comparing the results of multiple time periods. Finally, the result set of switchgear operation hidden danger identification is obtained.

[0094] S5 includes the following steps:

[0095] S501: Based on the switchgear operation hazard identification result set obtained in step S4, and combined with the change rate and evolution trend of nanoparticle characteristic parameters, the hazard development trend prediction method is used to predict the evolution direction of the operation hazard, and the hazard development trend prediction result set is obtained.

[0096] Based on the switchgear operation hazard identification result set, the process first performs dynamic analysis on each hazard type. Combining the change rate and evolution trend of nanoparticle characteristic parameters, the development direction of the hazard is predicted. The rate calculation is based on the increment and time difference of the nanoparticle characteristic parameters within a time window. The rate formula can be expressed as: ,in Indicates the first The characteristic values ​​of nanoparticles at time t. This represents the characteristic value at the previous time step. This represents a time interval. By analyzing the rate of change over multiple time periods, the evolution trend of potential hazards can be determined. This evolution trend can be fitted to historical data using a linear regression model to predict the direction of change in hazard types over a future period. The trend prediction formula is expressed as follows: ,in The predicted hazard level is represented by time t, a represents the regression coefficient, and b is the bias term. By comparing the current state with the predicted trend, it is determined whether the hazard is developing in the expected direction. Correction terms are introduced in the prediction process, such as changes in environmental factors or operating parameters. The correction coefficient can be calculated by the weighted average method. Finally, the predicted result set of hazard development trend is obtained.

[0097] S502: Based on the prediction result set of hidden danger development trend, a hierarchical early warning output method is adopted to generate corresponding level of operation hidden danger early warning information according to the prediction results and preset thresholds, thereby obtaining switchgear operation hidden danger early warning information.

[0098] Based on the predicted results of potential hazards, the execution process first compares the predicted results with preset hazard level thresholds. The threshold classification is obtained through statistical analysis of historical hazard occurrence data. Common preset levels can be divided into low, warning, and severe. Low level indicates that the hazard is developing slowly, warning level indicates that the hazard has a risk of further development, and severe level indicates that the hazard may cause serious failure. Different level thresholds are set according to different hazard types and their development trends. For example, for hazards caused by abnormal temperatures, the threshold setting may be determined based on historical temperature change data over a period of time, setting an upper limit for the rate of temperature change. If the rate exceeds this threshold, it is classified as a warning level. For hazards related to humidity changes, different threshold ranges may be set. If the humidity change exceeds a certain set value, a severe warning will be triggered. The predicted results are compared with these thresholds, and corresponding hazard warning information is generated based on different threshold ranges. Finally, all warning information is summarized and output to obtain the switchgear operation hazard warning information.

[0099] The switchgear operation hazard monitoring and early warning system based on nanoparticle detection includes the following modules: a nanoparticle sensing module, which is used to perform multi-parameter real-time detection of nanoparticles in the air inside the switchgear based on the electrical stress and thermal stress generated during the operation of the switchgear, and to perform data preprocessing and time series construction on the obtained nanoparticle particle size parameters, number concentration parameters and change characteristics to generate the original nanoparticle operation characteristic dataset.

[0100] The feature analysis and correction module is used to receive the original operating feature dataset of nanoparticles, perform time series feature analysis on nanoparticle parameters, extract trend features, fluctuation features and abrupt change features, and combine them with switchgear operating environment parameters to perform adaptive correction and correlation screening of features, generating a set of corrected electrical features parameters of nanoparticles.

[0101] The hazard warning and decision-making module receives the set of electrical-related modified feature parameters of nanoparticles, matches and analyzes them with the pre-built switchgear operation hazard feature model to identify the type and risk level of switchgear operation hazards, predicts the development trend of operation hazards, outputs corresponding level operation hazard warning information, and generates switchgear operation hazard warning information.

[0102] The nanoparticle sensing module includes:

[0103] The particle acquisition submodule is used to sample nanoparticles in the air inside the switchgear in real time based on the electrical and thermal stress generated during the operation of the switchgear, and to obtain the particle size parameters, number concentration parameters and distribution parameters of the nanoparticles.

[0104] Based on the electrical and thermal stresses generated during the operation of the switchgear, the particle acquisition submodule first analyzes the intensity and duration of these stresses. By monitoring the switchgear's operating status through sensors, it determines the variation patterns of these stresses. For example, if the duration of electrical stress exceeds a certain set threshold, the particle acquisition submodule initiates high-frequency sampling to capture changes in nanoparticles and record particle size, number concentration, and distribution parameters. Particle size refers to the diameter of the captured nanoparticles in the airflow; number concentration refers to the number of particles per unit volume; and distribution parameters describe the uniformity of particle distribution in the air. All parameters are time-stamped during sampling to ensure the timeliness and relevance of subsequent data. Through continuous sampling and data recording, the particle acquisition submodule obtains a large amount of operational data, providing raw data support for subsequent signal processing and timing modeling. Finally, it organizes and outputs the particle size, number concentration, and distribution parameters of the nanoparticles, forming an effective data acquisition set.

[0105] The signal preprocessing submodule is used to filter, denoise, and smooth the raw sampling signal acquired by the particle acquisition submodule in order to eliminate the influence of environmental interference and random noise.

[0106] First, the raw sampled signal acquired from the particle acquisition submodule is filtered using a bandpass filter to remove low-frequency and high-frequency noise. The filter design is based on the known electrical and thermal stress spectrum ranges, assuming that electrical stress-related noise frequencies are in the low-frequency band and thermal stress-related noise frequencies are in the high-frequency band. The filter cutoff frequency is adjusted based on experimental data and set to a specific range, for example, a low cutoff frequency of 1Hz and a high cutoff frequency of 100Hz. Next, denoising is performed using a wavelet transform algorithm. By selecting appropriate wavelet basis functions, random noise in the original signal is separated, and non-target signal components in the spectrum are removed. Finally, the processed signal is smoothed using a moving average algorithm, averaging the signal over multiple points within a time window to reduce high-frequency fluctuations and obtain a smoothed signal for subsequent analysis. The preprocessed signal is clearer and has lower errors, thus improving the accuracy of subsequent analysis. The final output is the filtered, denoised, and smoothed signal data.

[0107] The temporal construction submodule is used to continuously sample and organize the preprocessed nanoparticle parameters to construct temporal nanoparticle operational status data.

[0108] First, the nanoparticle parameters after signal preprocessing are time-series encoded. Specifically, within a continuous time window, data is time-calibrated according to the sampling frequency, and the corresponding particle parameter values, such as particle size and number concentration, are recorded at each time point. Assuming a 5-second time window is set within a certain sampling period, data is collected every 5 seconds, and all collected particle data are arranged chronologically to form continuous time-series data. Subsequently, the time-series construction submodule organizes this time-series data, calculates trend changes, and analyzes the parameter variation patterns within each time period. If the particle number concentration changes significantly in a short period, its trend is determined. If the trend is continuous and stable, it indicates that the particle concentration may be increasing. This data can provide important evidence for subsequent hazard prediction and alarm. After the time-series data is constructed, all data are arranged linearly over time, forming a complete time-series nanoparticle operational status dataset, ready for subsequent analysis and processing, ultimately generating a time-series dataset.

[0109] The data integration submodule is used to fuse and encapsulate time-series nanoparticle operational status data to form a dataset of original nanoparticle operational characteristics.

[0110] First, the time-series data on the operational status of nanoparticles are fused using a weighted average method. Data types are weighted according to importance and relevance, with important parameters such as particle size and number concentration receiving higher weights, while less important parameters such as particle distribution receive lower weights. For example, particle size and number concentration are weighted at 0.6 and 0.4 respectively, and particle distribution at 0.2. The weight coefficient for each data point is determined based on previous experimental data and data correlation analysis. By using weighted averaging, data points from multiple time windows are integrated to obtain more accurate data values. During the integration process, if a parameter is missing, interpolation is used to fill in the missing data. Linear interpolation is used to calculate a reasonable value for the missing position based on the values ​​of the preceding and following data points. After integration, the data is packaged into a standardized format to ensure its suitability for subsequent hazard analysis and processing, ultimately forming the original nanoparticle operational characteristic dataset, which is used for subsequent hazard identification and predictive analysis.

[0111] The feature analysis and correction module includes:

[0112] The trend extraction submodule is used to extract the trend features of nanoparticle parameters over time based on the original operational feature dataset of nanoparticles.

[0113] Based on the original operational feature dataset of nanoparticles, the trend extraction submodule first decomposes the time-series data formed by continuous sampling, extracting the particle size change sequence, quantity change sequence, and distribution change sequence separately, and aligning them according to a unified time scale. Then, it calculates the change slope for each type of parameter within multiple consecutive time windows, constructing the change slope by the difference between parameters at adjacent time points and the time interval. The change slope is obtained using an approximately linear calculation method and is used to describe the upward or downward trend of the parameter over time. When the slope is in a higher range, it is determined to be an upward trend; when it is in a lower range, it is determined to be a downward trend; and the middle range is considered a stable state. The interval division is set based on the statistical range of historical operational data. Then, the trend results within multiple time windows are compared, and the dominant trend is determined by the frequency of occurrence. For example, if the same change direction is shown in multiple consecutive windows, it is recorded as a continuous trend feature. During the execution process, for example, if the particle quantity parameter continuously increases in adjacent windows within a certain operational cycle, the trend calculation result is marked as monotonically increasing. Finally, the trend results of various parameters are summarized to form the trend feature extraction result.

[0114] The fluctuation analysis submodule is used to analyze the periodic fluctuation characteristics and abnormal fluctuation characteristics of nanoparticle parameters.

[0115] First, the time series of nanoparticle parameters is segmented, dividing the complete sequence into multiple equally long analysis segments. The mean and deviation of the data within each analysis segment are calculated. The deviation is obtained by the difference between a single point value and the mean within the segment. When the deviation amplitude is within the normal range, it is judged as periodic fluctuation. When it exceeds the preset fluctuation range, it is recorded as abnormal fluctuation. The fluctuation range is set with reference to the statistical results of historical stable operation periods, and is determined by comprehensively taking the deviation range of multiple operating cycles. Then, the periodic fluctuation is judged by comparing the time interval of the fluctuation peak in adjacent analysis segments. If the time interval is approximately consistent, it is marked as a periodic fluctuation feature. If the time interval has no obvious pattern and the deviation amplitude is large, it is classified as an abnormal fluctuation. For example, if the particle size parameter shows a large shift in a short period of time and does not correspond to the existing cycle, it is identified as an abnormal fluctuation record. Finally, the fluctuation analysis results are formed.

[0116] The mutation identification submodule is used to detect sudden changes in nanoparticle parameters and identify abnormal mutation characteristics.

[0117] First, differential calculations are performed on adjacent sampling points in the time series. The presence of sudden changes is determined by calculating the magnitude of changes between sampling points before and after the change. The magnitude of change is described by the relative rate of change, which is calculated as the ratio of the current change to the value at the previous moment. When the rate of change exceeds a pre-set mutation threshold range, the time point is identified as a mutation candidate point. The mutation threshold is set based on the upper limit of the normal fluctuation range. Then, multiple sampling points before and after the mutation candidate point are compared to determine whether the change is persistent. If the change quickly returns to the original level in a short period of time, it is marked as an instantaneous disturbance. If the change persists, it is recorded as an abnormal mutation feature. For example, if a quantitative parameter shows a significant jump in a very short period of time within a certain runtime segment and then remains in the new level range, the position is recorded as a mutation point. Finally, the mutation identification result is output.

[0118] The environmental correction submodule is used to adaptively compensate and dynamically correct the nanoparticle characteristic parameters by combining the temperature parameters, humidity parameters and background particulate matter concentration parameters in the operating environment of the switchgear.

[0119] First, temperature, humidity, and background particulate matter concentration parameters at the same time scale are acquired and standardized to ensure comparability of data with different dimensions. Then, the nanoparticle characteristic parameters are compared with the environmental parameters one by one. By calculating the proportional relationship between the changes in environmental parameters and the changes in nanoparticle parameters, a compensation coefficient is constructed. This compensation coefficient is set with reference to the corresponding relationship within the historical stable environmental range. For example, when the ambient temperature is in a high range, a corresponding correction amount is introduced for the particle number parameter. The correction amount is obtained by multiplying the original parameter and the compensation coefficient. Then, the correction result is superimposed or canceled with the original value to form the corrected parameter value. During the execution process, for example, when the humidity is in a high range, the particle size parameter is corrected downward, and finally, the dynamic correction of the nanoparticle characteristic parameters is completed.

[0120] The relevant screening submodule is used to use correlation analysis to screen parameters that are highly correlated with electrical operating conditions from the modified feature parameters, forming a set of modified electrical feature parameters for nanoparticles.

[0121] First, each correction parameter is matched with an electrical operating status indicator to construct a parameter comparison sequence. Then, a correlation calculation method is used to judge the correlation by comparing the consistency between the direction of parameter change and the direction of electrical status change. The correlation index can be described by the relative change synchronization rate, that is, the proportion of changes in the same direction. When the proportion is in a high range, it is judged as highly correlated; in a medium range, it is judged as generally correlated; and in a low range, it is judged as weakly correlated. The range division is determined based on the statistical results of historical operating samples. Then, all parameters are screened, and only parameters with a high correlation level are retained. For example, if a certain particle quantity parameter keeps synchronized with the change of electrical load in multiple operating cycles, it is retained as a valid parameter. Finally, the screened parameters are summarized to form a set of nanoparticle electrical correlation correction characteristic parameters.

[0122] The hazard warning and decision-making module includes:

[0123] The hazard matching submodule is used to match and analyze the feature parameters with a pre-built typical operation hazard feature model of switchgear based on the set of electrical-related modified feature parameters of nanoparticles, and obtain the hazard type matching results.

[0124] First, the input multidimensional feature parameters are decomposed, extracting particle number variation indicators, particle size distribution variation indicators, and related fluctuation indicators, and mapping them uniformly to a pre-defined feature dimension space. Then, a corresponding comparison sequence is constructed for each type of feature parameter. Matching analysis is completed by comparing the deviation of the current feature parameter with the reference parameter in the typical hidden danger feature model. During the comparison, a relative difference calculation method is used, dividing the difference between the current parameter and the model parameter by the baseline interval width of the model parameter to obtain the normalized matching deviation. When the deviation is in a lower interval, it is judged as highly similar; in the middle interval, it is judged as partially similar; and in the higher interval, it is judged as dissimilar. The division of each interval is set according to the parameter fluctuation range in historical operating samples. In specific execution, for example, if the particle number variation feature and the corresponding parameter in the poor contact type hidden danger model show similar variation amplitude in a certain period, then the parameter is marked as a matching item. Then, the matching results of multiple parameters are summarized, and the overall matching degree is obtained by weighted summation. The weight setting refers to the frequency of occurrence of different parameters in historical hidden danger samples. Finally, the hidden danger type matching result is output.

[0125] The risk assessment submodule is used to determine the development stage and risk level of potential hazards in switchgear operation based on the hazard type matching results.

[0126] First, the matching results are structured, using the matching degree values ​​corresponding to different hazard types as input parameters. The changes in matching degree are analyzed in conjunction with time-related information. By comparing the current matching degree with reference intervals of similar historical hazards at different development stages, the current stage is determined. The stage intervals are set based on the statistical distribution of the gradual changes in matching degree during hazard evolution. When the matching degree is in the initial interval, it is marked as the early stage; in the middle interval, it is marked as the development stage; and in the higher interval, it is marked as the aggravated stage. Then, risk level calculation is introduced based on the stage determination. By combining stage weights with matching degree weights, a comprehensive risk value is obtained. The comprehensive risk value is calculated using a simple linear superposition method. The weights are set with reference to the recorded proportions of different stages in historical operation. For example, if the matching degree continuously rises and enters the high interval within a certain operating cycle, the stage weights are adjusted accordingly. Finally, the risk level is determined based on the interval of the comprehensive risk value, and the corresponding risk determination result is output.

[0127] The trend prediction submodule is used to predict the development direction of potential operational hazards by combining the rate of change and evolution trend of the characteristic parameters of nanoparticles.

[0128] First, the rate of change of the characteristic parameters of nanoparticles over a continuous time period is calculated. The rate of change is obtained by the ratio of the parameter difference within adjacent time windows to the time span. The rate results of multiple parameters are then summarized and processed. Subsequently, the evolution direction extracted from the previous trend characteristics is combined to determine the sign and magnitude of the rate of change. When the rate is consistently positive and the magnitude is in a high range, it is recorded as an accelerated change state, and when it is in a low range, it is recorded as a slow change state. The division of each range is set based on the rate statistical range during normal operation. On this basis, the rate of change of multiple time periods is compared to determine whether it shows an increasing, decreasing, or fluctuating state. For example, if the particle parameter rate gradually increases over several consecutive periods, the evolution trend is marked as a continuously increasing direction. Then, the rate judgment result is combined with the trend direction to form the prediction result of the hidden danger development direction.

[0129] The early warning output submodule is used to generate corresponding level of operational hazard early warning information based on the risk level and prediction results, and output the switchgear operational hazard early warning information.

[0130] First, the risk level results and trend prediction results are compiled and matched. Different risk levels and different development directions are combined to form various early warning conditions. Then, the early warning levels are divided according to the early warning conditions. The level range is set with reference to the record distribution of different hidden danger states in historical operation. By combining the risk level value and the trend prediction label, the early warning level corresponding to the current state is determined. For example, if the risk level is in the medium range and the trend prediction result shows a continuous strengthening direction, the corresponding early warning level is marked as medium-high level. Then, the descriptive information associated with the early warning level is retrieved to form a complete early warning content structure. Finally, according to the preset data output format, the early warning level and corresponding content are encapsulated to output the switchgear operation hidden danger early warning information.

[0131] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring and early warning of potential operational hazards in switchgear based on nanoparticle detection, characterized in that, Includes the following steps: S1: Real-time sampling of nanoparticles in the air inside the switch cabinet; using a multi-parameter synchronous acquisition method for nanoparticles to obtain the particle size distribution parameters, quantity concentration parameters, and unit time change parameters of nanoparticles; and continuously recording them through time series sampling to obtain the original operating characteristic dataset of nanoparticles. S2: Based on the original operational feature dataset of the nanoparticles, the time series feature extraction method and the sliding time window analysis method are used to extract the trend features, fluctuation features and mutation features of the nanoparticle parameters, construct the time series feature expression of the nanoparticle operational state, and obtain the time series evolution feature parameter set of nanoparticles; S3: Based on the nanoparticle time-series evolution characteristic parameter set, combined with the switchgear operating environment parameters, including temperature parameters, humidity parameters and background particulate matter concentration parameters, the environmental impact adaptive correction method is used to dynamically compensate and correct the nanoparticle characteristic parameters to obtain the nanoparticle electrical related correction characteristic parameter set; S4: Based on the set of electrical-related modified feature parameters of the nanoparticles, the hidden danger pattern matching identification method is adopted to compare the modified feature parameters with the pre-established typical hidden danger feature model of the switchgear, and the operation hidden danger type and corresponding risk level are identified by feature similarity calculation and classification judgment method to obtain the switchgear operation hidden danger identification result set. S5: Based on the identified results set of potential hazards in the switchgear, and combined with the change rate and evolution trend of the nanoparticle characteristic parameters, a hazard development trend prediction method is used to predict the direction of hazard evolution, and corresponding level hazard warning information is generated through a graded warning output method to obtain hazard warning information for switchgear operation.

2. The method for monitoring and early warning of potential operational hazards in switchgear based on nanoparticle detection according to claim 1, characterized in that: S1 includes the following steps: S101: Based on the electrical and thermal stresses generated during the operation of the switchgear, nanoparticle detection units are deployed inside the switchgear and along the ventilation path to sample nanoparticles in the air inside the cabinet in real time, obtain the particle size parameters, number concentration parameters, and distribution parameters of the nanoparticles, and obtain the original monitoring dataset of nanoparticles. S102: Based on the original monitoring dataset of nanoparticles, the sampled data is filtered, denoised and smoothed using data preprocessing methods to eliminate random noise and unstable interference, and a preprocessed dataset of nanoparticles is obtained. S103: Based on the nanoparticle preprocessing dataset, the nanoparticle parameters are continuously recorded and organized using a time series sampling method to form time-series data for subsequent analysis, thus obtaining the original nanoparticle operational characteristic dataset.

3. The method for monitoring and early warning of potential operational hazards in switchgear based on nanoparticle detection according to claim 1, characterized in that: S2 includes the following steps: S201: Based on the original operational feature dataset of nanoparticles obtained in step S1, the trend characteristics of nanoparticle size and number concentration changing with time are analyzed using the time series feature extraction method to obtain the time series trend feature dataset of nanoparticles. S202: Based on the nanoparticle time-series trend feature dataset, the periodic fluctuations and abnormal fluctuations of nanoparticle parameters are extracted using the fluctuation feature analysis method to obtain the nanoparticle time-series fluctuation feature dataset. S203: Based on the nanoparticle temporal fluctuation feature dataset, the sudden change feature of nanoparticle parameters is identified by the mutation point detection method, forming a complete temporal feature expression, and obtaining the nanoparticle temporal evolution feature parameter set.

4. The method for monitoring and early warning of potential operational hazards in switchgear based on nanoparticle detection according to claim 1, characterized in that: S3 includes the following steps: S301: Based on the nanoparticle time-series evolution characteristic parameter set obtained in step S2, temperature parameters, humidity parameters and background particulate matter concentration parameters in the switch cabinet operating environment are collected simultaneously to form an environmental impact parameter dataset; S302: Based on the environmental impact parameter dataset, an environmental impact adaptive correction method is used to dynamically compensate and correct the nanoparticle characteristic parameters to obtain an environmentally corrected nanoparticle characteristic parameter set. S303: Based on the environmentally corrected set of nanoparticle characteristic parameters, a correlation analysis method is used to screen characteristic parameters that are highly correlated with the electrical operating state, thereby obtaining a set of nanoparticle electrical correlation corrected characteristic parameters.

5. The method for monitoring and early warning of potential operational hazards in switchgear based on nanoparticle detection according to claim 1, characterized in that: S4 includes the following steps: S401: Based on the set of electrical-related modified feature parameters of nanoparticles obtained in step S3, the feature parameters are matched with the pre-established typical hidden danger feature model of switchgear using the hidden danger pattern matching identification method to obtain a preliminary result set of hidden danger type matching. S402: Based on the preliminary result set of the hazard type matching, the feature similarity calculation method is used to accurately identify the hazard type, and a precise hazard type matching result set is obtained; S403: Based on the accurate matching result set of the hazard types, the risk level determination method is used to determine the development stage and risk level of the hazard, and the result set of hazard identification for switchgear operation is obtained.

6. The method for monitoring and early warning of potential operational hazards in switchgear based on nanoparticle detection according to claim 1, characterized in that: S5 includes the following steps: S501: Based on the switchgear operation hazard identification result set obtained in step S4, and combined with the change rate and evolution trend of nanoparticle characteristic parameters, the hazard development trend prediction method is used to predict the evolution direction of the operation hazard, and the hazard development trend prediction result set is obtained. S502: Based on the predicted result set of the hidden danger development trend, a hierarchical early warning output method is adopted to generate corresponding level of operation hidden danger early warning information according to the prediction results and preset thresholds, thereby obtaining switchgear operation hidden danger early warning information.

7. A switchgear operation hazard monitoring and early warning system based on nanoparticle detection, characterized in that, It includes the following modules: a nanoparticle sensing module, which is used to perform multi-parameter real-time detection of nanoparticles in the air inside the switch cabinet based on the electrical and thermal stress generated during the operation of the switch cabinet. The module performs data preprocessing and time series construction on the obtained nanoparticle particle size parameters, number concentration parameters and change characteristics to generate a dataset of original nanoparticle operating characteristics. The feature analysis and correction module is used to receive the original operating feature dataset of the nanoparticles, perform time series feature analysis on the nanoparticle parameters, extract the trend features, fluctuation features and mutation features, and combine the switch cabinet operating environment parameters to perform adaptive correction and correlation screening on the features, and generate a set of corrected electrical features parameters of nanoparticles. The hazard warning and decision-making module is used to receive the set of electrical-related modified feature parameters of the nanoparticles, match and analyze them with the pre-built switchgear operation hazard feature model to identify the type and risk level of switchgear operation hazards, predict the development trend of operation hazards, output corresponding level operation hazard warning information, and generate switchgear operation hazard warning information.

8. The switchgear operation hazard monitoring and early warning system based on nanoparticle detection according to claim 7, characterized in that: The nanoparticle sensing module includes: The particle acquisition submodule is used to sample nanoparticles in the air inside the switchgear in real time based on the electrical and thermal stress generated during the operation of the switchgear, and to obtain the particle size parameters, number concentration parameters and distribution parameters of the nanoparticles. The signal preprocessing submodule is used to filter, denoise, and smooth the raw sampling signal acquired by the particle acquisition submodule in order to eliminate environmental interference and random noise. The temporal construction submodule is used to continuously sample and organize the preprocessed nanoparticle parameters to construct temporal nanoparticle operational status data. The data integration submodule is used to fuse and encapsulate the time-seriesd nanoparticle operational status data to form the original operational feature dataset of the nanoparticles.

9. The switchgear operation hazard monitoring and early warning system based on nanoparticle detection according to claim 7, characterized in that: The feature analysis and correction module includes: The trend extraction submodule is used to extract the trend features of nanoparticle parameters changing over time based on the original operational feature dataset of the nanoparticles. The fluctuation analysis submodule is used to analyze the periodic fluctuation characteristics and abnormal fluctuation characteristics of nanoparticle parameters. The mutation identification submodule is used to detect sudden changes in nanoparticle parameters and identify abnormal mutation characteristics. The environmental correction submodule is used to adaptively compensate and dynamically correct the nanoparticle characteristic parameters by combining the temperature parameters, humidity parameters and background particulate matter concentration parameters in the operating environment of the switchgear. The relevant screening submodule is used to use correlation analysis methods to screen parameters that are highly correlated with the electrical operating state from the modified feature parameters, forming the set of electrical-related modified feature parameters of the nanoparticles.

10. The switchgear operation hazard monitoring and early warning system based on nanoparticle detection according to claim 7, characterized in that: The hazard early warning decision module includes: The hidden danger matching submodule is used to perform matching analysis between the feature parameters and the pre-built typical operation hidden danger feature model of switchgear based on the set of electrical-related modified feature parameters of the nanoparticles, so as to obtain the hidden danger type matching result; The risk assessment submodule is used to determine the development stage and risk level of the switchgear operation hazards based on the hazard type matching results. The trend prediction submodule is used to predict the development direction of potential operational hazards by combining the rate of change and evolution trend of the characteristic parameters of nanoparticles. The early warning output submodule is used to generate corresponding level of operational hazard early warning information based on the risk level and prediction results, and output the switchgear operational hazard early warning information.

Images